Leadscrew or ballscrew delta

DjDemonD

Okay, so I'm musing on making a leadscrew or ballscrew based delta printer with a twist, as its going to use an acrylic cylinder for a frame.

I am not doing this because I think it will be a better configuration than same an all-metal kossel XL, which I know works pretty damn well. It's more an exercise in building something unusual and interesting.

My question is on leadscrews and ballscrews.

1)Pitch, I understand that to get reasonable speed I am not going to be able to maximise resolution by having fine pitch screws, otherwise my motors just won't be able to turn them very fast. 8mm?
2) How much motor torque is needed to drive them leadscrew vs ballscrew? The mass on each tower is not going to be large, (carriage, rods, share of the mass of the effector)

Any thoughts? Anyone tried this already?

deckingman

I know nothing about Deltas but the ex engineer in me used to know a bit about lead screws and ball screws. In a nutshell, ball screws were invented as a low friction alternative to lead screws. That's all I'm prepared to say about your proposition - (other than best of luck Simon).

bot

You could try acquiring some of these: http://www.linengineering.com/products/stepper-motor-linear-actuators/ln-stepper-motor-linear-actuators/

I believe robotdigg sells something like that.

It might allow you to move faster, with the right pitch, as you wouldn't be spinning the entire length of the screw – just the nut. With the drawback of moving the motor on each carriage.

DjDemonD

Thanks Ian (I think :)).

DjDemonD

Interesting Bot, I think that might be a bit heavy duty, I'm looking at probably (as acrylic tubes are hard to get so I'm using a Biube fishtank) 30cm diameter and 45 cm height. So nema 17's and electronics top mounted with the intention that the base is open and can be placed onto whatever you want to print on, like David's scara using blue tape on the desk.

I have in mind to use mgn9 rails, mount the bearing blocks for the screws to the lower end of the rails and design carriages with the nut in them.

I also need to figure out how to precision drill the holes to mount the rails. I'm thinking either some sort of multi axis cnc or maybe just a paper template printed from cad drawings. I might be able to print a scaffold with holes marked on my large delta, assuming it is accurate enough and I'm not going to be introducing analogue copying error.

dc42

If the screws are 8mm lead then the steps/mm using 1.8deg motors @ x16 microstepping would be 200 * 16/8 = 400. That's already 4x higher than a typical belt-drive delta. Another way of looking at it is that it's like using a 4-tooth GT2 pulley (same 8mm movement/rev). So you certainly don't want a smaller lead than 8mm, and you shouldn't use 0.9deg motors. You will probably need 24V power to get reasonable speeds.

deckingman

Also consider that 400 micro steps per mm is 25 full steps, or 2.5 full steps per 0.1mm. So any odd value of layer height will rely on micro-stepping for positional accuracy which isn't generally best practice. Even numbers (0.2, 0.4 etc) would be fine though (well that is on a "normal" printer)

timcurtis67

You could also reduce the micro steps if you don't want 400 steps per mm. 1/4 step gets you back to matching what a delta belt moves.

DjDemonD

David - thanks, yes 24v was certainly on the list.
Timcurtis67 - It'll make a lovely noise at 1/4

But that is interesting, whether microstepping might be useful is uncertain, with leadscrews I doubt it, but ballscrews should offer very low friction, like the rails themselves.

toysrfun

@djdemond

For others reading this I’d run separate drivers through duex5 and 3000rpm servos and it will fly with fine pitch screws setup direct drive or 1:1. I have a CNC mill with 10mm pitch and very large servos that can run 1200ipm. Off top of head that is in 30,000mm/m and extremely fast I can’t imagine a 3D printer keeping up. At 10mm pitch I have 2.5 tpi and a 5mm pitch at approx 5tpi would potentially cut the speed in half. That’s still faster than you need so consider a 4mm or 2mm pitch. Ultimately you can go many of ways and use gear reduction if needed. The accuracy and positional feedback would be a huge improvement over belts if setup right. If using a course pitch to attempt industrial speeds you will have other challenges including z axis holding. I would in that case be sure the z axis servo had a built in off power brake. You can run steppers for extruders and servos for your axis. Presuming you want to spend $240 per axis. Personally I would do it and not waste my money again on steppers. Once you’ve changed you have a difficult time ever looking back. Everyone will say steppers are fine, blah, blah, and I’d be willing to bet it comes down to their own williness to spend the money on quality products. After all for most it’s a weekend hobby and nothing more. For others it’s the future and the sky is the limit. For myself I won’t even say what I’ve spent on this current project. Good thing I’m not married

Danal

I'd encourage you to use 9mm belts (instead of 6mm). They are quite close to half as stretchy as 6, and high quality 9mm work very well.

I do understand your desire for "something different", so if you must "Get Screwed"... Ballscrews all the way!!! Leadscrews are a "middle ground" or "compromise" and there is no reason to go away from belts and NOT go to the best of the best in your other option.

Danal

And... if you want to be REALLY different, have you considered this delta mechanical setup? Those are servos up top. Fast, beefy, servos.

https://youtu.be/Ec3Vf8Ip4Ic?t=17